Entrepreneur Launches exe.dev: A Cloud Infrastructure Built on Fundamentally Different Principles
A tech founder who co-established a successful startup is now launching a new cloud computing platform designed to address what he sees as systemic flaws in today's major cloud providers.
The announcement of exe.dev marks a significant pivot for the entrepreneur, who acknowledged that founding a second company while already running a thriving business venture surprised many of his peers in the startup community. However, he explained his decision stems from a straightforward motivation: his genuine enjoyment of working with computers and frustration with existing cloud infrastructure.
The Problem with Current Cloud Solutions
The founder articulated several fundamental issues he identifies with contemporary cloud providers. He contends that virtual machine architecture is structurally flawed, as current offerings tie computational resources directly to CPU and memory allocation. This prevents users from purchasing resources independently and managing multiple VMs flexibly on hardware they control.
Platform-as-a-Service offerings, he argued, create vendor lock-in by requiring developers to learn proprietary systems for each cloud provider, only to discover mid-project that standard computing tasks become prohibitively difficult due to undisclosed platform limitations.
Storage and Performance Challenges
The entrepreneur highlighted how cloud storage architecture, designed when hard drives dominated the market, remains poorly optimized for solid-state drives. Remote block devices introduced when seek times measured in milliseconds now impose unacceptable performance penalties on SSDs, which operate at microsecond speeds. He noted that configuring an EC2 instance for 200,000 input/output operations per second costs approximately $10,000 monthly, while a MacBook achieves 500,000 IOPS without such expense.
Networking and Cost Structure
Data egress pricing from major cloud providers costs roughly ten times more than traditional data center pricing. While hyperscalers offer improved rates for high-volume customers spending millions monthly, most projects operate on far smaller budgets, making cloud expenses prohibitive for practical applications.
Why Now? The AI Factor
The founder identified artificial intelligence and large language models as the catalyst for rebuilding cloud infrastructure. He explained that as AI agents become capable of writing code independently, software production will increase dramatically—a phenomenon economists describe as Jevons paradox. This expansion demands more accessible, efficient computing resources than existing cloud platforms provide.
AI agents themselves struggle with fundamental cloud abstraction limitations, consuming excessive context tokens attempting to navigate complex cloud APIs rather than solving user problems.
The exe.dev Solution
The new platform introduces a different architectural approach. Rather than selling individual virtual machines, exe.dev allows customers to purchase CPU and memory resources, then run as many VMs as desired on acquired hardware. The service includes integrated TLS and authentication proxies to prevent machines from being directly exposed to the internet.
Storage uses local NVMe drives with asynchronous block replication for redundancy. The platform operates data centers across multiple global regions, with anycast networking providing low-latency access to deployed applications for international users.
Planned features include static IP addresses and user-friendly interfaces for accessing automated historical disk snapshots. The founder indicated the team is simultaneously racking physical computers in data centers and reconsidering every layer of the software infrastructure stack.
The founder concluded by emphasizing his personal motivation: "I am building a cloud. One I actually want to use. I hope it is useful to you."